---
title: "Launch one LULA-2 adapter fine-tune job."
method: POST
path: "/v2/lula2/finetune"
tags: ["LULA"]
---

# Launch one LULA-2 adapter fine-tune job.

`POST /v2/lula2/finetune`

## Request

- Base URL: `https://api.omtx.ai` (the document declares no server; this is the origin it was published from)
- URL: `https://api.omtx.ai/v2/lula2/finetune`
- Auth: one of:
  - API key in header `x-api-key`
  - HTTP bearer

## Headers

- `Idempotency-Key` string, required

## Request body

- object: Public LULA-2 hosted fine-tune request.
  - `job_name` string: Human-readable job name
  - `csv_artifact_id` string, uuid, required: Uploaded UTF-8 .csv artifact with target_id, protein_sequence, smiles, and label columns; split is required only for train_test
  - `training_pair_count` integer, required: Exact accepted CSV row count; verified before training
  - `training_mode` 'train_test' | 'train_all': train_test uses split=train for optimization and split=test for holdout metrics; train_all optimizes on every accepted row
  - `base_derived_model_id` string, uuid, nullable: Optional owned ready derived model for continuation
  - `seed` integer: Deterministic training seed

## Response `200`

Successful Response

- unknown

## Other responses

- `202`: Job accepted. Poll /v2/jobs/{job_id} for status/result.

---

[API](https://skmtc.dev/omtx/apis/gateway.md) - [All operations](https://skmtc.dev/omtx/apis/gateway/llms.txt) - [OpenAPI document](https://skmtc.dev/omtx/apis/gateway/revisions/232715a75205?raw)